The LeptonX Suite

Three products.
One sovereign platform.

Every component runs entirely on-device — no cloud, no exposure, no compromise on privacy or clinical precision.

■ Patent Pending — U.S. Application No. 64/000,111

Product 01

LPMA

Local Private Medical Associate

Your complete medical intelligence hub. LPMA ingests clinical records via Epic MyChart and HL7/FHIR standards, builds a private vector database of your health history, and enables natural voice queries — all processed locally with zero cloud data transfer.

For the first time, you can ask "what did my oncologist say about my last scan?" and get a synthesized, accurate answer in seconds — from a system that has read every page of your medical record.

HIPAA HL7/FHIR Epic MyChart Zero Cloud
Request Early Access
LPMA medical intelligence
LPMA — Core Platform
Product 02

SarcomaClaw

Specialized Oncology Intelligence

A purpose-built RAG engine for sarcoma oncology. SarcomaClaw synthesizes patient-specific records with current treatment literature, IHC pathology data, and clinical trial information — achieving 94.6% benchmark accuracy on a rigorous 44-query evaluation suite (97.1% excluding a confirmed source data gap).

Unlike general-purpose medical AI, SarcomaClaw was built for a specific clinical challenge by someone navigating it in real life. That specificity is the difference between an answer and the right answer.

Current benchmark 94.6%
Labs100%
Treatment96%
View Full Benchmark
SarcomaClaw oncology
SarcomaClaw — Oncology AI
Product 03

SpeakClaw

Clinical Voice Assistant

Ask clinical questions naturally. Hear synthesized answers that draw from your complete medical record and current literature — with correct pronunciation of every physician name, medication, and clinical term.

The custom phoneme lexicon ensures "Dr. Aisling" is spoken correctly, "trabectedin" is pronounced with clinical authority, and every response opens with empathy — not "Right" or "So."

On-Device STT + TTS (Parakeet + Kokoro)
Custom Clinical Phoneme Lexicon
Streaming Low-Latency Response
Context-Aware Query Routing
Request Demo
SpeakClaw voice assistant
SpeakClaw — Voice Agent
Coming Soon — Product 04

Mobile Intelligence Container

A temporary, self-contained intelligence package generated by your LeptonX hub and exported to your mobile device for offline use — at appointments, in low-connectivity settings, or anywhere you need medical AI without a network.

No sync. No bidirectional complexity. Generated on demand, used locally, deleted after the session. Docker as the common denominator — platform-agnostic, secure, ephemeral.

Express Interest
Generate
Hub creates container

Selects quantized model, exports patient-specific encrypted vectors, transfers via local WiFi.

Use
Offline intelligence

Decrypts locally. Zero network dependency. Voice queries answered on-device.

Dispose
Clean deletion

Container deleted after session. No residual PHI on mobile device. Recreate fresh on demand.